Skip to main content
The honest answer to “will this work for my app?”

​
What Mini can do

Drive any iOS or Android app

Real builds running on cloud emulators. Taps, swipes, scrolls, types, waits, dismisses system dialogs, accepts permissions, recovers from unexpected screens. If a human can navigate it on a phone, Mini can.

Sign in with any credential flow

Email and password are the default. Mini also handles Sign in with Google via Minitap-managed accounts — pick one from the Shared by Minitap group in the profile picker. Custom auth (magic links, SMS codes, in-app email verification) works as long as you give Mini a profile with the credentials.

Author the suite for you

Once GitHub is connected, Mini reads your repository to draft the initial suite, add stories as features ship, retire stories whose features are gone, and rewrite criteria when screens change. Profiles and uploaded device files stay manual.

Build the app for you

Connect GitHub and Mini builds the app on its own infrastructure — no separate CI workflow needed for test builds. Or keep your existing workflow and upload the artifact yourself. Both paths feed the same run engine. See Providing app builds.

Run from anywhere you already work

CI, the dashboard, Slack, the CLI, and Cursor or Claude. A run triggered from Slack lands in the same Issues tab as a run triggered from CI.

Produce a fix prompt, not just a bug report

When something fails, Mini delivers the criterion that didn’t hold, a video of the moment of failure, and a fix prompt you can paste into Cursor or Claude. Device logs come along when they help explain what happened.

Triage from Slack or the dashboard

Acknowledge, mark as not-a-bug, mark as resolved — same actions, same state, regardless of surface. The Issues tab is the canonical view, but most triage happens in the Slack thread where the issue first surfaced.

Override criticality per issue

Mini infers criticality at runtime, but you have the final word. Override it per issue when the default doesn’t match your business reality.

Catch regressions you didn't write tests for

Suggestions surface things that look wrong but aren’t tied to any acceptance criterion — visual regressions, copy changes, broken navigation. They show up in the Suggestions tab in the dashboard.

Navigate OS surfaces

Mini can open the Settings app, accept notification prompts, toggle permissions, switch to another app (like a browser for OAuth) and return via deep link, control network conditions to test under slow connectivity, and pull device logs into the run report.

​
What’s on the way

Scheduled runs

Cron-style scheduling. Today the workaround is a GitHub Action that pushes a tag on a schedule.

Hardware features

Cloud emulators handle most flows. Support for camera, biometrics, and push notifications is coming.

Retry from Slack

The @mini retry command is stubbed. Today, use the dashboard’s Re-run button or re-trigger from CI.

​
What Mini can’t do today

Real limits in the current product. Some are temporary, some are permanent.
  • Microphone. Mini can’t speak into the mic or pipe audio in. Voice-driven flows are out of scope.
  • Camera. Mini can’t inject a specific image into the camera. QR scanning or document photographing needs a different approach.
  • Biometrics. Touch ID / Face ID prompts can be dismissed, but Mini can’t enroll or authenticate a real biometric.
  • Bluetooth, NFC, motion sensors. Anything requiring a physical device beyond touch input.
  • External browsers. Mini stays in your app. If sign-in opens Safari/Chrome and bounces back via deep link, that works. Living in the browser doesn’t.
  • Email and SMS verification. Mini can read the last code from a fixed email address on the profile (for sign-up flows). It can’t open a real inbox, click an arbitrary link, or read SMS.
  • VPN configuration, Bluetooth pairing, and similar multi-step system flows that go beyond your app’s scope.
  • Networking shims. Mini drives the actual app talking to your actual backend. No intercepted network calls, no injected responses, no mocked services.
  • Time travel. Mini can’t set the device clock. “What happens on January 1st” needs to be triggered on January 1st.
  • Concurrent users. One run, one user. Multi-user scenarios need separate sequential runs.
  • Web apps in a desktop browser. The product targets native mobile. Mini can drive a PWA installed on a phone, but not a website opened on a laptop.
  • Tablet-specific layouts. Mini runs on phone-sized emulators. iPad and large-screen tablet UIs aren’t covered.
  • Watch and TV apps. Out of scope.

​
What’s not on the roadmap

To keep expectations clear:
  • Unit tests and integration tests of your code. Use your existing tooling.
  • Load testing and performance benchmarks. Mini runs a handful of devices, not a fleet.
  • Security testing and penetration testing.
  • A general-purpose mobile RPA platform. Mini is built for QA, not for automating real user workflows.
If something on your wishlist isn’t listed here, ask. The roadmap moves fast.